Bought some gravel yesterday for this purpose. Tried it out with cotton and gravel only. As expected it drained well. Then tried some sand. It occurred to me that the sand would mix with the gravel and make it very difficult to separate these out when I disassembled the apparatus later. Not only was this true, but the filtration was a disappointment.
Arg!
It looks like the sand will clog up the paper, which is intended to keep them separated. Furthermore, the clogging prevents much water from reaching the end point of the filtration.
The trick is to allow the sand to filter out the most of the crud from the water, while keeping it from getting down into the cotton, where it gets clogged.
Small rocks should be at the border between sand and bigger rocks below. The rocks should get gradually bigger. Hopefully, the small rocks will stop the sand seepage before it can clog up the cotton and/or paper.
I found in earlier experiments that paper worked better as a filter than cotton. Cotton should be the last boundary then, because you don't want anything impeding the flow of water. Small rocks should form the barrier between the cotton and the upper layers then.
Here's a scheme to try: Cotton paper mix at the last layer before exit. On top of that, place the smallest rocks and gradually increase the size of the rocks to the midrange, then make them smaller again as you approach the sand layer. The sand layer is next to last, with a layer of larger rocks on top of it. Why? To keep the sand form hollowing out where the water flow is the greatest.
Update: ( 10 am )
This post will be updated as I go along today in my lab. As I'm writing, I just got back from an errand, which meant the lab was interrupted briefly. Just before going on the errand, I had a slight mishap with my apparatus, which caused a big mess that I had to clean up. It's a little after 10 am local time, so I am about to go back into laboratory.
Just before the mess, I had been fooling around with the rocks and sand only to find that the sand will filter down through the rocks anyway. Whatever I do, the sand wants to defeat my purpose. So, I'm thinking up some ideas on what to do next. Maybe some hard barriers that will catch the sand and kept the two separated. Perhaps some aluminum foil with holes in it, which will allow water through, but hopefully not much sand.
Update ( 11:10 am ):
Well, this is getting a bit frustrating. Sand is very hard to keep out of where you don't want it to be. An idea occurred to me to use some microwave bowl covers with holes in them. These will allow water to go through, and does appear to stop the sand a little. Perhaps these can be stacked on top of each other because they are just about as wide in diameter as a 2 liter cola bottle.
Update ( 1 pm ):
This sand will not drain worth a flip. I've got a cola bottle end acting like a funnel with the sand filling it up a few inches and it won't even drain through the big hole in the cola bottle. What the heck? My dogs are barking and I'm a gonna take a break. This one's kicking my ass.
By the way, I'm taking a break while drinking a bottle of water. Bwah hah hah hah!
Update ( 3:40 pm ):
The day has gone by so fast. Yet, I'm not at a solution yet. In order to deal with the sand issue, I decided to try another way from those that I've seen on the YouTube. Instead of having the water go straight through the layers, I'll have it go up aways, then drain through holes on the side. That's sort of like a septic tank. Solids, meaning sand in this case, will go to the bottom, while relatively clearer water will go out the sides.
So far this hasn't produced dramatic results, but this is all rather crude anyway. It looks like it needs to settle for a lot longer than what I'm giving it time to.
If I were to tweak this a bit, I may get some good results.